Bug 390791

Carlos reports:

- Works fine when running the first 92 tests (i.e. ./TestRunner -r 10
`head -n 92 testlist`). Adding one more crashes the system.
- - Works fine running the last 95 tests (i.e. ./TestRunner -r 10 `tail
- -n 95 testlist`). Adding one more crashes the system.

Have you tried trimming from both ends; deleting both the first few and last few
tests.

Also, any chance you can get a screen-shot (even a camera :-) or text of the crash?

Phil's managing to build frysk on ppc now so we might be able to repo.

(In reply to comment #2)
> Have you tried trimming from both ends; deleting both the first few and last few
> tests.

No. I'll try that.

> Also, any chance you can get a screen-shot (even a camera :-) or text of the
crash?

No. I'm using a remote machine via ssh and it just reboots.

> Phil's managing to build frysk on ppc now so we might be able to repo.

Great! I hope his hardware is different than mine, so we can rule out any kind
of hardware incompatibilities. I'm building on a dual 970FX.

I have a F8 PPC4 machine that the test fail on, but do not produce rebooting
issues. I do not have the exact hardware specs for this machine, as it is
remote. The test result, as follows, while undesirable does not cause rebooting.
Perhaps there is another cause for the hardware fail? (ie previous tests run in
the suite?)

-bash-3.2$ ./TestRunner frysk.bindir.TestFhd
Running testHpdPid(frysk.bindir.TestFhd) ...ERROR
  frysk.expunit.TimeoutException: Timeout of 5 expired
Running testHpdCommand(frysk.bindir.TestFhd) ...PASS
Running testHpdCore(frysk.bindir.TestFhd) ...ERROR
  frysk.expunit.TimeoutException: Timeout of 5 expired

Time: 10.282
There were 2 errors:
1) testHpdPid(frysk.bindir.TestFhd)frysk.expunit.TimeoutException: Timeout of 5
expired
   at frysk.expunit.Expect.expectMilliseconds(Expect.java:241)
   at frysk.expunit.Expect.expect(Expect.java:277)
   at frysk.expunit.Expect.expect(Expect.java:297)
   at frysk.expunit.Expect.expect(Expect.java:317)
   at frysk.bindir.TestFhd.testHpdPid(TestFhd.java:75)
   at frysk.junit.Runner.runCases(Runner.java:207)
   at frysk.junit.Runner.runArchCases(Runner.java:241)
   at frysk.junit.Runner.runTestCases(Runner.java:512)
   at TestRunner.main(TestRunner.java:61)
2) testHpdCore(frysk.bindir.TestFhd)frysk.expunit.TimeoutException: Timeout of 5
expired
   at frysk.expunit.Expect.expectMilliseconds(Expect.java:241)
   at frysk.expunit.Expect.expect(Expect.java:277)
   at frysk.expunit.Expect.expect(Expect.java:297)
   at frysk.expunit.Expect.expect(Expect.java:317)
   at frysk.bindir.TestFhd.testHpdCore(TestFhd.java:94)
   at frysk.junit.Runner.runCases(Runner.java:207)
   at frysk.junit.Runner.runArchCases(Runner.java:241)
   at frysk.junit.Runner.runTestCases(Runner.java:512)
   at TestRunner.main(TestRunner.java:61)

FAILURES!!!
Tests run: 3,  Failures: 0,  Errors: 2

Failed after run #0
-bash-3.2$ 

Phil

The standalone test runs. The problem happens when the entire testsuite runs, or
when I run the commands in comment #1. Try those, please.

Regards,

Carlos

Carlos,

Can you try to get the console output from the remote crashing machine?  With
out it I'm not sure how much further progress is possible.

Unfortunately, there's no console output. Just a black screen and then, after
some time, the booting procedures. 

Couldn't you reproduce that?

Also, I was discussing with Jose Flavio here and we think this might be caused
by a memory leak in utrace/ptrace, since it takes a lot of tests until the crash
happens. What do you think?

Regards,

Carlos

doing some bz cleanup. RHEL 5.10 is very limited release, so closing this as will not fix.

-Steve